With the development in digital imaging devices, they have been capable of capturing images with high spatial resolution at relatively low cost. At the same time, in the aim of size reduction and low cost for cameras, an imaging element used in the cameras have been downsized. However, when an object moving in images is captured in dark environments, there is a problem of quality deterioration, which is reduction of a signal-to-noise (S/N) ratio, in capturing images having a high spatial resolution.
This problem is caused by the following reason. Since the number of pixels in an imaging element is increased or a size of the imaging element itself is reduced, a light-receiving area of each pixel is reduced. Thereby, the amount of received light is inadequate, which causes noises and eventually deteriorates image quality. Even if electrical signal amplification is performed after receiving light, noises are also amplified together. As a result, a S/N ratio is not improved.
Moreover, if an exposure period is extended to increase the amount of received light, motion blur is caused due to motion of an object. Therefore, clear images are not obtained. Furthermore, in order to increase the amount of received light, it has been also considered to enlarge an optical system such as a lens to increase an amount of collected light. This causes, however, a problem of increasing a size of an overall camera and eventually increasing a cost.
Among conventional apparatuses which generate dark and moving images, there is disclosed a technology of performing capturing a plurality of times (refer to Patent Reference 1, for example). FIG. 1 is a flowchart of processing performed by a conventional image generation apparatus according to Patent Reference 1.
As shown in FIG. 1, in Step 2201, test capturing is performed. Thereby, a direction and a magnitude of motion of an object are determined to decide how many times divided capturing processes are to be performed. In Step 2202, capturing processes are performed the decided times. In Step 2203, based on motion vector information, specific regions in respective captured frame images are shifted to average values of the frame images. Thereby, an eventual captured image is generated.    [Patent Reference 1] Japanese Unexamined Patent Application No. 2003-259184